Category:Online Argumentation

From Online-Partizipation
Jump to navigation Jump to search

Definition

Online argumentation is the exchange of arguments using Internet technology. In the context of online participation, the term argument includes formal elements, such as premise-conclusion structures, as well as informal aspects, such as stories and examples. In online participation processes, online argumentation is frequently a very important part of the execution phase.

Theories

Online argumentation in online participation processes is based on two lines of theories. On the one hand there is argumentation theory. This line of theory is about logical reasoning, it provides formal definitions of how an argument is composed, how arguments can relate to each other and what conclusions can be drawn from a set of arguments. Key disciplines involved in work on formal argumentation and argumentation theory are philosophy, (theoretical) computer science and mathematics. On the other hand there is deliberation. Deliberation refers to the ability to exchange information and opinions and to find agreements. Key disciplines involved in work on deliberation are political science and communication science. Deliberation is often considered to be a vital element in democratic processes.

The Behavior of Participants in Online Argumentation

The behavior of participants in online argumentation systems greatly depends on the platform design. For example, Esau et al. (2017) compare between different types of news platforms that differ in terms of platform design (a news forum, news websites, and Facebook news pages) and find out that deliberation (measured as rationality, reciprocity, respect, and constructiveness) differs significantly between platforms. The study shows that the news forum yields the most argumentation and respectful debate. While user comments on news websites are only slightly less deliberative, Facebook comments perform poorly in terms of deliberative quality. Having regard to argumentation, the results show that whereas a majority of comments in the news forum (72 percent) and news websites (56 percent) addressed at least one argument, only a third of the Facebook comments (33 percent) provided arguments. However, comments left on news websites and on Facebook show particularly high levels of reciprocity among users.

Methods for Analyzing the Results of Online Argumentation

As of now, most online participation processes utilize forum-based platforms on which users can enter unstructured texts. Automated analysis of the content of those texts is currently a research challange. There are two levels of automated analysis that are of interest in this context: A fine-granular analysis that revolves around the extraction of arguments and a more coarse-granular analysis that covers a broad overview of the discussion topics. The former is part of the research field argument mining and the latter is called topic extraction.

In argument mining, three classification tasks are commonly performed one after another: (i) argument identification; (ii) argument classification; and (iii) argument linking. The first step, argument identification, deals with the separation of non-argumentative text from argumentative text content as a binary classification task. Separation of argumentative text provides a brief overview over previous works on argument identification. Afterwards, argument classification deals with the extraction of argument components from an argumentative text. To this end it is necessary to specify an argument model. A basic argument model comprising claims and premises and extensions thereof is called claim-premise family. Claims are usually defined as controversial statements that are either true or false. Premises are reasons that either support or attack a claim. Please refer to this list of argument mining papers with models from the claim-premise family for more information on this. For German online participation processes, introduces a new argument model tailored for that specific use case. With argument linking as the last step of an argument mining system, relations between different argument components are automatically identified.

For topic extraction as the more coarse-granular analysis, simple keyword-based approaches can be used to extract the most frequent terms in an online participation process. With topic modeling techniques, such as Latent Dirichlet Allocation (LDA), it is possible to identify multiple discussion topics. Additionally, there are a lot of automated analysis tasks, e.g., the identification of expressed emotions.

Once Arguments have been properly identified, their interrelation can be displaye by argumentation maps. Formal analysis of argumentation maps is done mainly by the calculation of so called 'extensions' of a given snapshot of the discussion. The origins go back to the seminal work by Dung (1995). Many extending models and analytical methods have been proposed in the past, including Verification in Attack-incomplete Argumentation Frameworks and Verification in Argument-incomplete Argumentation Frameworks by Daniel Neugebauer and Hilmar Schadrack. See their Ph.D. Projects Formal Analysis of Online-Argumentation and Modeling Agents of Online Debates for more information.

Systems for Online Argumentation

There is a large number of systems available that support online argumentation. Those systems can be classified as follows:

Subcategories

This category has the following 5 subcategories, out of 5 total.

A

D

F

N

P

Pages in category "Online Argumentation"

The following 185 pages are in this category, out of 185 total.